Molecular Phylogenetics and Evolution 41 (2006) 345–354 www.elsevier.com/locate/ympev Pinniped phylogeny and a new hypothesis for their origin and dispersal Ulfur Arnason a,¤, Anette Gullberg a, Axel Janke a, Morgan Kullberg a, Niles Lehman b, Evgeny A. Petrov c, Risto Väinölä d a Department of Cell and Organism Biology, Division of Evolutionary Molecular Systematics, University of Lund, S-223 62 Lund, Sweden b Department of Chemistry, Portland State University, Portland, OR 97207, USA c The Eastern Siberian ScientiWc and Fishery Production Centre (VOSTSIBRYBCENTR), Ulan-Ude, Buryatia 670034, Russia d Finnish Museum of Natural History, P.O. Box 17, FI-00014 University of Helsinki, Finland Received 10 January 2006; revised 11 May 2006; accepted 19 May 2006 Available online 26 May 2006 Abstract The relationships and the zoogeography of the three extant pinniped families, Otariidae (sea lions and fur seals), Odobenidae (one extant species, the walrus), and Phocidae (true seals), have been contentious. Here, we address these topics in a molecular study that includes all extant species of true seals and sea lions, four fur seals and the walrus. Contrary to prevailing morphological views the analy- ses conclusively showed monophyletic Pinnipedia with a basal split between Otarioidea (Otariidae + Odobenidae) and Phocidae. The northern fur seal was the sister to all remaining otariids and neither sea lions nor arctocephaline fur seals were recognized as monophy- letic entities. The basal Phocidae split between Monachinae (monk seals and southern true seals) and Phocinae (northern true seals) was strongly supported. The phylogeny of the Phocinae suggests that the ancestors of Cystophora (hooded seal) and the Phocini (e.g. harp seal, ringed seal) adapted to Arctic conditions and ice-breeding before 12 MYA (million years ago) as supported by the white natal coat of these lineages. The origin of the endemic Caspian and Baikal seals was dated well before the onset of major Pleistocene glaciations. The current Wndings, together with recent advances in pinniped paleontology, allow the proposal of a new hypothesis for pinniped origin and early dispersal. The hypothesis posits that pinnipeds originated on the North American continent with early otarioid and otariid diver- gences taking place in the northeast PaciWc and those of the phocids in coastal areas of southeast N America for later dispersal to colder environments in the N Atlantic and the Arctic Basin, and in Antarctic waters. © 2006 Elsevier Inc. All rights reserved. Keywords: Pinnipedia; Phylogeny; Mitogenomics; Evolution; Biogeography 1. Introduction peds had originated in the Arctic Basin, with the otarioids subsequently colonizing the N PaciWc and the phocids the The topic of pinniped biogeography was probably Wrst N Atlantic, before dispersing to other areas. addressed by Sclater (1897), who postulated origin of the Davies (1958) had assumed monophyletic Pinnipedia in group in southern oceans. This view, which had been sec- his study. That notion was gradually replaced, however, by onded by von Boetticher (1934), was formally questioned a morphologically based view that favored pinniped by Davies (1958), who advocated pinniped origin in the diphyly with the otarioids sharing ancestry with the ursids Holarctic. Based on the working hypothesis that “The and the phocids with the mustelids. This phylogenetic pinnipeds are, and always have been, generally tied to a understanding also led to a revised pinniped biogeography, cold-water environment”, Davies concluded that the pinni- the essence of which was that early otarioid evolution had taken place in coastal areas of the N PaciWc and that of the phocids in European waters. * Corresponding author. Fax: +46 46147874. Sarich (1969) challenged the notion of pinniped diphyly in E-mail address: [email protected] (U. Arnason). an immunological study that unequivocally showed mono- 1055-7903/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2006.05.022 346 U. Arnason et al. / Molecular Phylogenetics and Evolution 41 (2006) 345–354 phyletic Pinnipedia. This phylogeny has been recognized in Table 1 essentially all later molecular analyses, irrespective of Taxa included in the phylogenetic analyses whether they have been based on nuclear data (de Jong and Pinnipedia Goodman, 1982; Arnason and Widegren, 1986), complete Odobenidae, walruses, one recent species mitochondrial (mt) genes (Arnason et al., 1995; Ledje and Odobenus rosmarus, walrus (AJ428576) Arnason, 1996a,b) or, more recently, on complete mt Otariidae, sea lions, fur seals genomes (Arnason et al., 2002; Arnason and Janke, 2002; Callorhinae Davis et al., 2004; Delisle and Strobeck, 2005). The molecular Callorhinus ursinus, northern fur seal (AM181016) recognition of pinniped monophyly gradually led to recon- Otariinae/Arctocephalinae sideration of the morphological understanding favoring pin- Otaria byronia, southern sea liona niped diphyly. However, these newer morphological studies Eumetopias jubatus, northern sea lion (AJ428578) (e.g. Wyss and Flynn, 1993; Berta and Wyss, 1994) did not Zalophus californianus, Californian sea lion (AM181017) Arctocephalus pusillus, South African fur seal (AM181018) support the traditionally accepted sister group relationship Phocarctos hookeri, Hooker’s sea lion (AM181019) between Otariidae and Odobenidae (DOtarioidea), but Neophoca cinerea, Australian sea lion (AM181020) rather advocated a sister group relationship between Phoci- Arctocephalus townsendi, Guadalupe fur seal (AM181021) dae and Odobenidae ( DPhocomorpha) to the exclusion of A. forsteri, New Zealand fur seal (AF513820) a Otariidae. This latter relationship was assumed as the basis A. australis, South American fur seal of a recent proposal for pinniped biogeography (Deméré Phocidae, true seals et al., 2002). More recently, Fyler et al. (2005) addressed some Monachinae, monk seals and southern true seals aspects of Monachinae biogeography. Some conclusions of Monachini this paper are inconsistent, however, with recent advances in Monachus monachus, Mediterranean monk seala phocid paleontology (e.g. Koretsky, 2001; Koretsky and M. schauinslandi, Hawaiian monk seal (AM181022) Holec, 2002; Koretsky and Sanders, 2002). Miroungini In the current study, we reconsider pinniped evolution Mirounga leonina, southern elephant seal (AM181023) a and biogeography using more extensive molecular data M. angustirostris, northern elephant seal than previously available, and by taking into account pale- Lobodontini a ontological evidence that has previously not been placed in Ommatophoca rossi, Ross seal a phylogenetic context of this kind. The analyses include 21 Lobodon carcinophagus, crabeater seal (AM181024) Leptonychotes weddelli, Weddell seal (AM181025) newly sequenced and 10 previously published pinniped mt Hydrurga leptonyx, leopard seal (AM181026) genomes. The sampling includes all recent species of true Phocinae, northern true seals seals and sea lions, four fur seals and the walrus. The analy- Erignathini sis followed the now common practice of joining the Erignathus barbatus, bearded seal (AM181027) sequences of 12 mitochondrial (mt) protein-coding genes of Cystophorini each species into one long sequence. This approach, intro- Cystophora cristata, hooded seal (AM181028) duced by Arnason and Johnsson (1992), expanded on by Phocini Janke et al. (1994) and extensively discussed by Paton and Histriophocina Baker (2006), rests on the demonstration that, while indi- Histriophoca fasciata, ribbon seal (AM181029) vidual genes may yield diVerent trees, longer genes and in Pagophilus groenlandicus, harp seal (AM181030) particular the concatenated sequences of all genes result in Phocina trees that show consistency with well-supported non-molec- Phoca largha, largha seal (AM181031) ular hypotheses. P. vitulina, harbor seal, Icelandic specimen (X63726) The initial phase of the study aimed at establishing pin- P. vitulina, harbor seal, Baltic Sea specimen (AM181032) niped relationships. Thereafter the phylogenetic results, in P. caspica, Caspian seal (AM181033) P. sibirica, Baikal seal (AM181034) conjunction with molecular datings and new paleontologi- P. sibirica, Baikal seal (AM181035) cal and climatological Wndings, were used to elaborate the P. hispida, ringed seal, Baltic Sea specimen (AM181036) new hypothesis for pinniped biogeography. Halichoerus grypus (Phoca grypa), grey seal (X72004) 2. Materials and methods Ursus americanus, black bear (NC003426) U. maritimus, polar bear (AJ428577) Mustela vison, American minka Table 1 lists the 39 species included in the study and the Gulo gulo, wolverinea accession numbers of their complete mt genomes (when Vulpes vulpes, red fox (AM181037) applicable). The accession numbers of the new genomes are Canis familiaris, dog (NC002008.4) shown in bold. The mt genome of the ringed seal, Phoca Felis catus, cat (U20753) Acinonyx jubatus, cheetah (AY463959) hispida, was sequenced using cloned restriction fragments Equus asinus, donkey (X97337) from enriched mtDNA preparations. The preparation of Rhinoceros unicornis, Indian rhinoceros (X97336) enriched mtDNA followed the procedure brieXy described in Arnason et al. (1991). Remaining genomes were PCR U. Arnason et al. / Molecular Phylogenetics and Evolution 41 (2006) 345–354 347 Table 1 (continued) 3. Results and discussion Pinnipeds are arranged according to a scheme that is essentially consistent with the current phylogenetic Wndings. Traditional Otariinae (sea lions) and